Development of an algorithm for processing “electronic nose” data on piezosensors when analyzing blood samples without sample preparation: a pilot study

The first results of blood analysis of patients of different profile departments without sample preparation using portable “electronic nose” on piezosensors are presented. During half a year on the basis of the clinical laboratory of the regional hospital the blood of patients was analyzed in parallel by traditional and sensor methods. The influence of conditions (room temperature, repetition rate, nature of modifiers of piezosensor electrodes) on the repeatability of sensor array signals was considered. Effective approaches and algorithms for processing multidimensional data of the piezosensor array for detection of the profile of volatile compounds (VOCs) of blood of small volume (not more than 0.5 ml) are proposed. The use of type II distilled water samples is effective as an internal standard when analyzing blood samples without sample preparation in a laboratory setting. Blood samples from 250 patients were analyzed. It was found that the set of piezosensors reliably distinguishes samples with pronounced pathologies of inflammation, oncology, serious problems in kidney function, the highest level of stress (surgery; traffic accident with injuries incompatible with life; burns). Other pathologies are also recorded with the proposed parameter, but its value depends on the individual state of the patient, the presence of comorbidities, the achievement of compensation, the degree of severity of negative processes on admission (for example, type 2 diabetes mellitus). The difference of VOC profiles for samples with different significant pathologies with component-by-component analysis is the subject of the next communication. The method of blood volatile compounds detection (measurement mode, repetition rate) is optimized and simple efficient algorithms for sensor data array processing are proposed.